yolov8在train时，出现box_loss、cls_loss、dfl_loss为nan，Box(P R mAP50 mAP50-95)为0的解决办法

python3.10，pytorch1.13，显卡1650 ，   yolov8版本    __version__ = '8.0.117'

yolov8n在train时，出现box_loss、cls_loss、dfl_loss为nan

model.train设置amp=False

model.train(data='VOC.yaml', epochs=100, imgsz=640, device=[0], batch=4, workers=0, amp=False)

 修改后 box_loss、cls_loss、dfl_loss显示数值，不是nan

 

Box(P R mAP50 mAP50-95)为0  的问题

解决办法：

修改

ultralytics/yolo/cfg/default.yaml     第49行   half  为   False 

half: False  # use half precision (FP16)

 

并注释掉ultralytics/yolo/engine/validator.py   中第102行代码   # self.args.half = self.device.type != 'cpu'     将self.args.half的值设置为False

self.training = trainer is not None
        if self.training:
            self.device = trainer.device
            self.data = trainer.data
            model = trainer.ema.ema or trainer.model

            # self.args.half = self.device.type != 'cpu'  # force FP16 val during training

            self.model = model
            self.loss = torch.zeros_like(trainer.loss_items, device=trainer.device)
            self.args.plots = trainer.stopper.poss